J O I N T C E N T E RAEI-BROOKINGS JOINT CENTER FOR REGULATORY STUDIES

The Future of Nuclear Power in the United States: Economic and Regulatory Challenges*

Paul L. Joskow†

Working Paper 06-25

December 2006

* This paper is based on a presentation that I made at the conference “Is Nuclear Power a Solution to Global Warming and Rising Energy Prices?” sponsored by the American Enterprise Institute in Washington, D.C. on October 6, 2006. This paper relies heavily on research performed in conjunction with the MIT Interdisciplinary Study The Future of Nuclear Power (MIT 2003), subsequent research that I have performed in more recent developments, and on the analysis contained in the International Energy Agency's World Energy Outlook 2006, Chapter 13 (IEA 2006) for which I was an advisor. The interpretations and conclusions contained in this paper are my own. † Elizabeth and James Killian Professor of Economics at the Massachusetts Institute of Technology (MIT) and Director, MIT Center for Energy and Environmental Policy Research.

J O I N T C E N T E R AEI-BROOKINGS JOINT CENTER FOR REGULATORY STUDIES

In order to promote public understanding of the impact of regulations on consumers, business, and government, the American Enterprise Institute and the Brookings Institution established the AEI-Brookings Joint Center for Regulatory Studies. The Joint Center’s primary purpose is to hold lawmakers and regulators more accountable by providing thoughtful, objective analysis of relevant laws and regulations. Over the past three decades, AEI and Brookings have generated an impressive body of research on regulation. The Joint Center builds on this solid foundation, evaluating the economic impact of laws and regulations and offering constructive suggestions for reforms to enhance productivity and welfare. The views expressed in Joint Center publications are those of the authors and do not necessarily reflect the views of the Joint Center.

ROBERT W. HAHN Executive Director

ROBERT E. LITAN Director

COUNCIL OF ACADEMIC ADVISERS

KENNETH J. ARROW Stanford University

MAUREEN L. CROPPERUniversity of Maryland

JOHN D. GRAHAM Pardee RAND Graduate School

PHILIP K. HOWARD Common Good

PAUL L. JOSKOW Massachusetts Institute of Technology

DONALD KENNEDY Stanford University

ROGER G. NOLL Stanford University

PETER PASSELL Milken Institute

RICHARD SCHMALENSEEMassachusetts Institute of Technology

ROBERT N. STAVINS Harvard University

CASS R. SUNSTEIN University of Chicago

W. KIP VISCUSI Vanderbilt University

All AEI-Brookings Joint Center publications can be found at www.aei-brookings.org

© 2006 by the author. All rights reserved.

Executive Summary

This paper examines the economic and regulatory challenges that must be faced by potential investors in new nuclear power plants in the United States. The historical development of the existing fleet of over 100 nuclear plants and their recent performance history are discussed. The pattern of re-licensing of existing plants and the implications for the role of the extended operation of the existing fleet in the overall electricity supply portfolio over the next 50 years is examined. The economic competitiveness of investments in new nuclear power plants compared to investments in alternative base load technologies is discussed under a variety of assumptions about construction costs, fuel costs, competitive and economic regulatory environments and various levels of carbon emissions prices affected competing fossil-fueled technologies. The paper then turns to a discussion of federal government efforts to facilitate investment in new nuclear power plants. These include streamlined licensing procedures and financial incentive provided by the Energy Policy Act of 2005. The regulatory and financial changes make investment in new nuclear power plants significantly more attractive to investors. However, in the longer run a combination of lower and more stable construction costs combined with carbon emissions charges are likely to be necessary to make investments in new nuclear plants significantly more attractive than investments in pulverized coal plants. Unresolved waste disposal policies and local opposition to new nuclear plants are likely to represent barriers to investment in new nuclear power plants in some areas of the country. First mover plants that can benefit from federal financial incentives are most likely to be built in states that continue to regulate generating plants based on cost-of-service principles, transferring construction cost and operating performance risks to consumers, and where there is room on existing sites to build additional nuclear capacity.

1

The Future of Nuclear Power in the United States: Economic and Regulatory Challenges

Paul L. Joskow

1. Introduction and Overview

Nuclear power now accounts for almost 20% of the electricity produced in the United

States. However, the last nuclear power plant completed in the U.S. entered commercial

operation in 1994 and no new nuclear plants have received construction permits since 1979. If

the existing fleet of nuclear plants were to run only until the end of their initial license periods,

the supply of electricity from nuclear power would begin to decline in about 2015 and reach zero

in about 2030. Under current economic conditions, these plants would be replaced primarily

with a mix of coal and natural-gas fired plants. However, a large fraction of the nuclear plants

now operating have received, applied for, or intend to apply for extensions on their operating

permits that would allow them to continue operating for an additional 20 years. If all existing

plants are re-licensed this will extend the date when nuclear electricity supplies begin to decline

until about 2035 and fall to zero in about 2050, absent construction of any new nuclear power

plants. Even if all existing nuclear plants are re-licensed for 20 addition years of operation, the

fraction of electricity produced with nuclear power in the U.S. will decline monotonically over

the coming decades if investment in new nuclear generating plants is not forthcoming.

Additional electricity supplies required to balance supply and demand in the next 25 years will

come primarily from coal and gas-fired power plants given current cost expectations and in the

absence of a "price" being placed on CO2 emissions.

The decision of most nuclear plant operators to apply for extensions of their initial

operating licenses reflects several factors. First, the performance of the existing fleet of nuclear

plants has improved dramatically over time in terms of capacity factors, operating costs, and

safety indicators. Second, the price of alternative fuels and the wholesale market price of

electricity have increased dramatically. Third, the incremental capital costs required to extend

the life of existing plants to meet license renewal requirements is typically small compared to the

total cost of a new coal or natural gas combined cycle (CCGT) plant. On a going forward basis,

life-extensions for existing plants are now very cost-effective and, for merchant plant owners,

quite profitable. These developments also favorably impact the relative economic evaluation of

2

investments in new nuclear plants compared to base load alternatives. However, on a going

forward basis private investors in new nuclear plants must expect to recover both their operating

costs and the much higher capital costs of building a new plant from revenues earned from

electricity produced by the plant. 1 The necessary revenues must come either from sales of

power at market prices in states that have adopted competitive market models or through

regulated retail prices determined through the regulatory process in states where generation

investments are still subject to regulation.

It is worth noting that two actual and potential institutional changes have taken place

since the existing nuclear fleet was built that will effect the economic evaluations of potential

private investors in new nuclear plants. First, in many regions of the country the electric power

sector has been restructured and the generating segment deregulated (Joskow 2006a). In these

regions, investments in new nuclear plants would be on a merchant basis and face standard

market risks and rewards associated with uncertain construction costs, operating performance,

and changes in market conditions. Even in states where generation investments are still

regulated, long-term commitments for new generating capacity must often be put out for

competitive bids and both long-term power supply contracts and generating plants built by

vertically integrated utilities will be subject to performance incentives that will shift more cost

and performance risk to investors than in the past. These changes work against capital intensive

generating technologies like nuclear power because they increase financing costs. Second, the

prospect of mandatory controls on CO2 emissions makes it possible that in the not too distant

future, CO2 emissions will incur a price either in the form of a "green tax" or, more likely,

payments for (or opportunity costs of allocated) CO2 emissions permits under a cap and trade

system as is the case in Europe. Placing a significant price on CO2 emissions would make

investments in nuclear power more attractive in both competitive and cost-based regulated

environments.

2. The Existing Nuclear Fleet

There are 104 commercial nuclear power plants operating in the U.S today. Eleven

nuclear plants were shut down during the 1990s, in several case, for economic reasons, prior to

1 The economic calculus may be different for state-owned firms and I will not discuss such firm here.

3

the end of their licensed operating lives. However, no operating nuclear plants have been shut

down since 1998. The development of the existing fleet of nuclear plants confronted numerous

problems along the way that continue to affect business perceptions about potential investments

in new nuclear plants. These problems included: lengthy licensing processes, large construction

cost overruns, long construction periods, poor availability, high operation and maintenance costs,

the need for early replacement of steam generators and other major pieces of equipment, public

opposition to construction in several regions of the country, and an accident at the Three Mile

Island (TMI) nuclear plant in 1979 that had an adverse effect on public perceptions of nuclear

power safety. This accident also led to increased construction costs for nuclear plants then under

construction due to post-TMI regulatory delays and plant design changes. Moreover, the federal

government has yet to deliver on its promise to take procession of spent fuel and other nuclear

waste and develop interim and permanent storage facilities for it. Several utilities completing

nuclear power plants during the 1980s and early 1990s faced serious financial difficulties as state

regulators sought to “disallow” portions of their plants’ construction costs from the regulated

prices for electricity they were permitted to charge. At least one utility went bankrupt and others

came close.

Moreover, during the late 1980s and 1990s, natural gas prices fell dramatically and price

forecasts made during this period for future natural gas and coal prices indicated that real

(controlling for general inflation) natural gas prices would rise slowly and real coal prices would

continue to fall slowly of the next 25 years (EIA AEO 1998, page 2). As a result, investment in

natural gas combined-cycle gas turbine technology (CCGT) was perceived as being a lower cost,

lower financial risk, and more politically acceptable base load generation alternative to both coal

and nuclear generating technology, especially as competitive wholesale electricity markets began

to emerge. Between 1997 and 2005 over 200 GW on new “on grid” generating capacity was

added in the U.S. (Joskow 2006a), most of it fueled by natural gas. No new nuclear plants have

been ordered in the U.S. since 1978. Orders for 124 nuclear units were eventually cancelled.

Nearly ten years ago, the 1998 EIA Annual Energy Outlook forecasted the gradual demise of

nuclear power, projecting that 24 nuclear plants would retire before the end of their license

expiration dates and that 65 plants would retire before 2020 (EIA AEO 1998, pp. 54-56).

However, the performance of the existing fleet of plants has improved dramatically over

time and the economics of both life-extension and new investment have changed significantly in

4

the last few years. As displayed in Figure 1, real non-fuel operation and maintenance costs have

fallen (on average) from about 20 mills/kWh in 1992 to 12.5 mills/kWh in 2005 (in constant

$2003) and total direct real operating costs have fallen from over 27 mills/kWh to less than 17

mills/kWh during this time period.2 As displayed in Figure 2, plant capacity factors have

increased from less that 60% in the late 1980s to close to 90% in the last five years.3 Various

safety indicia have improved considerably as well.4 These improvements in the performance of

the existing fleet of plants, combined with dramatic increases in natural gas prices and wholesale

electricity market prices in the last couple of years, have significantly changed perceptions about

the remaining economic lives of the existing fleet from what they were only a few years ago. As

I have already noted, there have been no plants retirements since 1998 and most of the existing

fleet has received, applied for or expects to apply for 20-year extensions on their licenses to

bring their licensed operating periods to 60 years. As of November 2006, 47 plants had received

license extensions, 9 license extension applications are pending, and another 25 have indicated

their intension to apply for license extensions.5 This will extend the life of the existing fleet of

nuclear plants considerably, with significant plant retirements now not beginning until about

2035 and the existing fleet closing down completely by about 2050 (assuming no further life

extensions). In the process of making investments to extend plant lives, some units are being

“uprated” modestly and EIA forecasts that this process will create about 3.2 Gw of additional

nuclear capacity from the existing fleet. Another 1.3 Gw of capacity is associated with the

refurbishment and return to service of TVA’s Brown’s Ferry 1 nuclear unit in 2007.

2 These numbers do not include all of the costs associated with these plants. Various owner’s costs including regulatory compliance costs, insurance, training, and general corporate overheads are not reflected in these numbers as far as I can tell. 3 There is some selection bias in these figures since when poorly performing nuclear plants close before their design lives their poor capacity factors disappear from the data. There is a close relationship between improvements in capacity factors and improvements in "variable" operating costs, since the latter are driven primarily by whether a plant to in commercial service for a period of time rather than by how many kWhs it generates during this time period. 4 Annual Report 2005, Institute of Nuclear Power Operation (INPO), http://www.nei.org/documents/WANO_Performance_Indicators_2005.pdf, accessed December 1, 2006. 5 Nuclear Energy Institute, http://www.nei.org/index.asp?catnum=2&catid=343, accessed November 27, 2006.

5

3. Investment in New Nuclear Plants

The changing regulatory and competitive environment The improved performance of the existing fleet of nuclear plants, higher fossil fuel

prices, new plant designs that vendors claim will result in lower and more predictable

construction costs, a streamlined construction and operating license process, the prospect of

future mandatory CO2 emission constraints, and federal financial incentives created by the

Energy Policy Act of 2005, together have had favorable effects on the economic attractiveness of

investments in new nuclear plants. In the end, whether or not a company invests in a new

nuclear plant will be a commercial decision that turns on the (risk adjusted) economic

attractiveness (cost and profitability) of an investment in nuclear power generating capacity

compared to investments in alternative generating technologies that can produce equivalent

quantities of electricity.

The economic environment in which nuclear investments will be considered by

electricity producers has also changed dramatically since the first fleet of nuclear plants was built

in the United States. Until the late 1990s most of the nuclear power plants were built by

investor-owned utilities which were subject to cost of service regulation and had de facto

exclusive geographic service territories where they charged customers regulated prices

determined based on cost of service principles that evolved during the 20th century. Regulation

insulated these firms from competition from lower-cost generating companies and typically

allowed them to shift a large fraction of any construction cost overruns and poor operating

performance to consumers through regulated prices. Consumers arguably got the benefits of

what was then perceived to be an efficient industry structure built around the assumption that

electricity would be supplies most efficiently to a specific geographic area by vertically

integrated monopolies subject to price, cost and service quality regulation.

This is not to say that utilities building nuclear plants under traditional regulatory

arrangement during the development of the existing fleet faced no risk from poor performance,

since some regulatory agencies disallowed what they perceived to be imprudent costs and others

introduced incentive mechanisms to stimulate better operating performance (Joskow and

Schmalensee 1986). However, it is fairly clear from the record during this time period when

6

companies were frequently applying for rate increases, that consumers bore a large share of the

costs associated with construction costs overruns, poor operating performance, and changing

economic conditions. This regulatory environment made it possible for many utilities to finance

nuclear plants using a relatively low cost of capital. Since nuclear plants were and are more

capital intensive than pulverized coal or CCGT technology (absent carbon capture and

sequestration faxcilities), this regulatory environment favored nuclear power by reducing the cost

of capital, other costs held constant.

This regulatory and competitive environment has changed in many areas on the U.S. In

most of the Northeast, portions of the Midwest and in Texas, competitive wholesale electricity

markets have been created and competing merchant generators are free to participate as suppliers

in these competitive wholesale markets (Joskow 2006a). Wholesale prices for power have been

deregulated and as transition arrangements come to an end, retail consumers pay prices that

reflect competitive wholesale market prices not prices determined by cost-of-service regulation.

(Joskow 2006a). In the first instance, investors bear all construction cost, operating

performance, fuel price, and wholesale power market risks, as well as residual regulatory risks

associated with a restructuring and deregulation process that is still a work in progress.

These developments in the competitive and regulatory environment change the financial

risks associated with nuclear power plant investments in a number of ways. Development costs

associated with licensing, site studies and engineering studies, are not recoverable through a

regulatory process but must be recovered from market revenues when and if a plant is completed

and is operating. Construction cost overruns, major equipment replacement costs, variations in

availability and variations in market price will all be risks that the plant investor will bear. Of

course, investor can seek to hedge some or all of these risks through contractual arrangements

with vendors, construction companies, and forward contracts with market intermediaries and

retail consumers. The 1600 Mw nuclear power plant being built by AREVA-Siemens in Finland

has hedging arrangements of this kind on both the construction cost and energy sale sides (IEA

2006, p. 352). However, it is unlikely that the long-term contractual arrangements that the owner

of the new nuclear plant in Finland has with large buyers of the power supporting the project

could be replicated widely in those areas of the U.S. that rely on competitive wholesale and retail

electricity markets. Let me note clearly as well, that in a competitive market environment it is

the investors who must be convinced about expected construction costs and plant performance

7

parameters and the ultimate cost and profitability of a potential nuclear power plant project in

order to get them to invest. Lobbying to convince government official that investment in new

nuclear plants represents the lowest cost alternative in regions that have adopted competitive

models is not particularly productive.

There are many states in the U.S., however, which have not implemented comprehensive

restructuring and deregulation programs for electricity generation (Joskow 2006). Retail prices

continue to be regulated, utilities have de facto geographic monopoly service territories where

consumers are charged for electricity based on regulated prices, and utilities can assume that

prudently incurred generation costs will be passed through in regulated retail prices. In a

regulated environment utilities do have to spend a lot of time convincing regulators that their

proposed investments would be less costly than alternatives in order to get permission to proceed

with the project and to avail themselves of the protections of cost-of-service regulation. In these

states lobbying by nuclear suppliers to convince government officials to accept optimistic cost

projections may have some value. However, the regulatory environment that prevailed when the

existing fleet of nuclear plants was built is also changing and is likely to change further as

proposals by utilities to build new regulated “cost-of-service” generation projects increase.

First, many state commissions now require utilities to benchmark proposals to build their

own generating plants subject to cost of service regulation against proposals from independent

power producers to supply them with equivalent quantities of power under long term contracts.

These contracts in turn typically have incentive provisions that place construction cost and

operating performance risks on the owner of the plant rather than on the utility buying the power

and ultimately on the utility's customers paying regulated retail prices that include contractual

payments under purchased power contracts. Accordingly, in order to get permission from a state

regulatory agency to build a new plant under cost of service regulation, whether nuclear or non-

nuclear, the regulatory agency will have to be convinced that the proposed project will be less

costly for consumers than alternatives based on market benchmarks. Moreover, I believe that it

is likely that when regulators do allow vertically integrated utilities to build additional plants

under cost of service regulation that they will include incentive or performance-based regulation

provisions that are not unlike those in long-term purchased power contracts. Specifically, I

expect that they will place more of the risk of construction cost overruns and poor operating

performance on the utility and less on the utility's retail customers than was the case when the

8

existing nuclear fleet was under construction. These incentive arrangements will increase the

risks faced by utilities compared to the old regulatory regime and increase the cost of capital

properly attributed to the project.

Life-cycle cost comparisons This discussion leads naturally to a discussion of what we know about the expected future

life-cycle costs of new nuclear power plants compared to pulverized coal and CCGT technology,

the two primary baseload alternatives to nuclear, at least over the next 25 years if there is no

price places on CO2 emissions (AEI AEO 2006). Will potential investors view nuclear

investments as being the lowest cost or the most profitable base load generation alternative?

Here I draw primarily on work done in connection with the MIT Future of Nuclear Power Study

(2003) and more recent work published by the International Energy Agency (IEA 2006, Chapter

13) in 2006.

There are several primary variables that determine the expected life-cycle costs of

alternative generating technologies with a given generating capacity. These are (a) the

“overnight” construction cost, (b) the construction period and associated cash flows, (c)

financing costs and any associated cash flow constraints, (d) corporate income tax rules, (e) fuel

costs and their assumed real rate of escalation over time, (f) non-fuel operation and maintenance

expenses, (g) replacement capital expenditures during the life of the plant, (g) the life-cycle

capacity factor of the plant, and (h) the economic life of the plant. The future values of all of

these variables that will be realized in practice are necessarily uncertain at the time development

of a project begins. Let me say a few things about several of these cost drivers.

The overnight construction cost of a generating plant is a value that is often

misunderstood. It is the hypothetical cost of a generating plant if it could be built instantly

("overnight"). The number does not reflect inflation, the costs of construction financing, or the

length of time that it takes to build the plant and associated cash flows. Accordingly, the

overnight cost value must be combined with assumptions about the construction period, the

profile of cash flows during construction, real input price escalation during construction,

financing costs, taxes, insurance and other variables to arrive at the total costs of building the

plant up to the point when it enters commercial operation. Once the plant is completed,

financing costs, cash flow constraints associated with debt repayment schedules, taxes,

9

equipment replacement costs, fuel, operation and maintenance costs, insurance, and the assumed

capacity factor (production of electricity per unit of capacity), are then combined to yield a

pattern of expected post-construction “break-even” cash flows. The pre-construction and post-

construction cash flows are combined and then discounted back to the present (or the time the

project under consideration would proceed with the licensing and developing process) using the

assumed post-tax weighted average cost of capital relevant to the project. This present value of

future cash flows is then “levelized” over the life-cycle output of the plant in constant dollars to

yield a levelized real life-cycle cost of producing a kWh of electricity for each technology. One

can think of this as an annualized real life-cycle cost per kWh for each technology or, in a market

context, as the real life-cycle price per kWh that the plant would have to realize in a competitive

market to make it a break-even investment (just profitable enough to cover its total costs,

including the opportunity cost of equity invested in the plants). Different assumptions about the

overnight construction cost, financing costs, income tax rates and associated accounting

procedures, capacity factors, fuel price escalation and a few other variables can have very

significant effect on the results.

There has been substantial controversy over the appropriate assumptions for the

overnight construction costs for nuclear plants. Unlike, pulverized coal and CCGT plants, there

is no recent construction cost experience for nuclear plants in the U.S. or Europe. The nuclear

plant vendors have advertised very low overnight construction cost numbers based on their own

non-transparent engineering cost calculations. These projections should be viewed with some

skepticism both because they are not backed up with firm financial commitments and because

the nuclear industry has a poor record of forecasting construction costs (IEA 2006, page 372).

Moreover, these cost estimates appear to exclude certain “owner’s costs” such as engineering

costs, site acquisition costs, insurance costs, spare parts inventory costs, training costs, licensing

and other regulatory costs. I consider most of the recent vendor cost estimates to be implausibly

low.

The MIT study made use of actual construction cost experience for recent nuclear plants

completed in Japan and South Korea (MIT 2003, Appendix to Chapter 5). I have supplemented

that analysis with more recent information on the reported contractual cost of the nuclear plant

being built by AREVA-Siemens for TVO in Finland (which appears to exclude owner’s costs

and some have argued does not cover the total costs that AEVA-Siemens will actually incur), the

10

reported cost quoted for the plant being built for Electricté de France (EDF),6 also supplied by

AREVA, to confirm what is a good base case overnight cost estimate for a new nuclear plant.

The number used as the base case overnight cost estimate in the MIT study was $2,000/kW in

constant $2002 or about $2,300/kW in constant $2006. I believe that this is still good base case

overnight construction cost estimate, recognizing that there is considerable uncertainty about

what actual construction costs will turn out to be. This base case overnight cost number is much

lower than the construction costs experience for nuclear plants completed in the U.S. during the

1980s and early 1990s but much higher than vendor cost forecasts.

The MIT study also made the base case assumption that new nuclear plants could be

completed in five years, as does the more recent IEA study (IEA 2006, p. 365). This is below

the construction periods experienced by every major nuclear power country except Japan (IEA

2006, page 373) and, I consider it to be reasonable, though perhaps on the optimistic side. The

base case in the MIT study also assumed an 85% life-cycle capacity factor, though alternative

cases were considered as well. The 2006 IEA study makes similar assumptions about life-cycle

nuclear capacity factors. This value is lower than that assumed in most vendor presentations

(e.g. 90+%), but is significantly higher than the global life-cycle capacity factors realized

historically by the world's nuclear power plants.7 In this regard, it should be noted that low

capacity factors that may be associated with the “break-in” of new plants right after they are

completed yield larger economic penalties in a present-value calculation than do low capacity

factors late in the life of a plant. Other assumptions used in the MIT study can be found in the

Appendix to Chapter 5 of the MIT study (MIT 2003) and for the IEA study in Chapter 13 (IEA

2006, p. 365).

Table 1 reports estimates of the real levelized costs of nuclear plants, pulverized coal

plants, and CCGT plants under a variety of different assumptions.8 The first column reflects the

assumption that the plant will be built as a stand-alone merchant project with a 40-year life. The

6 An EDF Press release issued May 4, 2006 quotes a construction cost number of 3.3 billion euros for a 1600 MWe plant or over $2,500/kW at current exchange rates. AREVA-Siemens committed to build the equivalent plant in Finland for 3 billion euros, but construction delays have likely increased the actual costs of building it. AREVA has been criticized for not recovering the engineering and development costs it incurred to develop the 1600 Mw EPR in the revenues it expects to earn from these first two projects, but I think that these criticisms are unfair. 7 A few countries, including Finland, have life-cycle capacity factors that are slightly higher than 90%, but this has not been the norm. 8 Note that only about 3% of the electricity supplied in the U.S. relies on oil-fired generating plants. They are used primarily for peaking (low load facto) purposes.

11

competitive market environment in which merchant plants would be built is reflected in

assumptions about the cost of equity and debt capital, the debt-equity ratio, debt repayment

constraints, taxes, etc. The second column performs similar calculations for a regulated plant

financed by a utility on its balance sheet using a lower cost of capital than the merchant plant and

a longer amortization period. Implicit in this case is the assumption that consumers bear most of

the construction cost and operating performance risks associated with the project. In both cases,

we initially assume that the cost of capital for nuclear plants is higher than those for coal and gas

CCGT plants due to cost, regulatory and other uncertainties.

It is evident in Table 1 that with the base case assumption about the overnight

construction cost of a nuclear plant, the nuclear plant is not competitive with pulverized coal and

is only competitive with CCGT generating capacity when expected real life-cycle natural gas

prices are high (about $7.60/Mcf expressed in $2006 general price levels). For nuclear to be

competitive with coal, overnight construction costs must fall by about 25% from the base case

assumptions and financing costs must be equivalent to those for coal and CCGT plants. Thus, if

nuclear is to be competitive with fossil-fuel alternatives, especially coal, construction costs must

fall significantly from our base case assumptions and there must be no cost of capital premium

required to finance nuclear plants. Even here, the cost saving from building a nuclear plant

rather than a coal plant is quite small.

Table 2 reports the same results for the merchant case but includes prices for carbon

dioxide (CO2) emissions. The CO2 prices vary from about $13/ton to about $55/ton.9 These

prices could result from the imposition of a mandatory cap and trade program as in the EU or

from a tax on CO2 emissions. Obviously, CO2 prices burden fossil-fueled plants and make

nuclear plants relatively more attractive economically. At somewhere between $25/ton and

$50/ton CO2, nuclear becomes economical in almost all cases unless gas prices return to low

levels and even if significant reductions from the base case overnight construction cost numbers

cannot be realized. It is interesting to note that it is likely that CO2 prices within this range over

the next 50 years would be necessary to stabilize CO2 emissions by 2050 and to make carbon

capture and sequestration an economical option for using coal to generate electricity (MIT Future

of Coal Study, in process).

9 Note that the carbon (C) prices utilized in the MIT study have been expressed in terms of equivalent CO2 prices here using an adjustment factor of 3.67 to reflect the differences in atomic weights of C and CO2.

12

The IEA released its World Energy Outlook 2006 in November 2006 (IEA 2006). It

includes a detailed analysis of the future of nuclear power globally in Chapter 13. The focus on

nuclear power in this study was stimulated by concerns about CO2 emissions and the fact that

nuclear power is an electricity supply option that generates electricity without producing

significant quantities of CO2. The results of the IEA study of the expected costs of nuclear

power and associated uncertainties are reasonably similar to those in the MIT study. The

expected cost per kWh from new nuclear plants varies from about 5.0 cents/kWh to about 8.0

cents/kWh in $2005, depending on assumptions about discount rates (cost of financing) and

overnight construction costs (IEA, 2006, p. 368). The range of real levelized costs from reported

in Table 1 above, after adjusting the numbers to bring them to $2005 price levels for

comparability with the IEA study, are about 4.0 cents/kWh to about 7.5 cents/kWh. The IEA

numbers do not include the very low construction cost case ($1500/kW) in the MIT study, so

they are otherwise quite comparable. For nuclear to be competitive with coal both a low

discount rate and low construction costs must be realized if there is no price place on CO2

emissions. I do not believe that the low discount rate assumption that the IEA seems to focus on

subsequently and in the summary of the chapter on nuclear power is consistent with investment

by private firms that do not receive subsidies and pay income and property taxes in either a

competitive environment or in a regulatory environment with significant construction cost and

operating performance sharing incentive arrangements.

The IEA study also examines the impact of CO2 prices on the relative economics of

nuclear vs. fossil fuel alternative generating technologies. From the figures in the IEA report it

looks like the impacts of CO2 prices are qualitatively similar to those in the MIT study. With a

price between $30 and $50/ton of CO2, nuclear would be competitive with coal even with

relatively high overnight construction costs and high financing costs (discount rates).

The figures in Table 1 and Table 2 and those in the IEA study, combined with uncertainty

about licensing and permit approval costs and success rates, are consistent with the lack of much

commercial interest by investors in new nuclear power plants in the U.S. until relatively

recently.10 In my view, the primary case for investment in new nuclear plants is that they will

displace CO2 producing generating technologies, especially coal. The combination of efforts to

10 Unlike several countries in Europe there has never been a formal ban on building new nuclear power plants in the U.S,

13

streamline the licensing process without sacrificing safety, financial incentives included in the

Energy Policy Act of 2005, and a growing awareness of the implications of possible future U.S.

policies to place a price on CO2 emissions (e.g. passage of some variant of the McCain-

Leiberman Act and continuing diffusion of individual state CO2 mitigation policies), have been

the primary contributors to a renewal of interest in investments in new nuclear plants in the U.S.

However, absent significant CO2 mitigation benefits, the economic advantages of nuclear power

over coal (either PC or IGCC meeting EPA air quality standards for criteria pollutants) in the

U.S. are, at best, too small to justify other baggage that comes with nuclear power: safety risks,

proliferation risks, and waste disposal problems. I turn to the recent efforts to stimulate

investment in new nuclear plants in the U.S. in the next section.

4. Institutional Changes and Financial Incentives

Streamlined NRC licensing process

The Nuclear Regulatory Commission (NRC) has adopted new regulatory procedures to

reduce the costs and delays associated with licensing new commercial nuclear plants. First, the

NRC now has a process in place to certify the safety of Standard Plant Design Specifications

before the associated plants are included in a specific construction and operating license

application for a new plant. Once a Standard Plant Design has received safety certification from

the NRC, plant developers who choose to adopt the certified plant design as approved do not (in

theory) have to go through a new plant design safety review for each specific plant seeking a

construction and operating license.11 As of November 2006, three plant designs have been

certified through this process, one plant design is near the completion of the certification process,

and five plant designs are in a pre-application review process at the NRC (including AREVA's

EPR, the first model of which is under construction in Finland).12

Second, the NRC has created a new early site permitting process that permits companies

to pre-license sites where they may ultimately choose to build a new nuclear plant, but can do so

before the decision is made to go forward with a plant. Thus site certification can proceed and

be separated from the construction and operating license process. As of November 2006 four

11 This means that as a practical matter the U.S. will abandon the historical approach which did not rely on standardized plant designs. The U.S. has learned at least something from the French.

14

applications have been made under this early site permitting process.13 However, it is not

necessary for a developer to go through this process. A developer can choose instead to go

directly to the Combined Construction Permit and Operating License (COL) process which will

include site certification.

The new COL process is the third innovation that the NRC has introduced. Under the

regulatory process in place when the existing fleet of plants was built, each plant had to first go

through a construction license process that involved a safety review of both the plant and the site.

Once a construction permit was issued the developer could then proceed to build the plant.

When the plant was completed the owner/operator then had to apply for a separate operating

license. There were significant delays at each stage of this process, including delays resulting

from appeals to the federal courts and resulting extensions of the evidentiary regulatory hearing

process. The NRC has now created a streamlined licensing approach in which a certified plant

design (and an early site permit if one has been obtained) can go directly into a Combined

Construction and Operating License (COL) process. Once the COL is granted the developer can

proceed to build the plant. When it is completed the NRC then would perform a series of

inspections, verifications, tests and analysis to confirm that the plant was built in accordance

with all COL license criteria. A second regulatory proceeding to review a separate application

for an operating license would not be required.

In theory, this new regulatory process should reduce licensing costs and delays.

However, the COL process has not been tested as no proposed plants have yet gone through it

from beginning to end. Accordingly, how well it will work in practice, and how it will be

affected by hostile intervenors in the licensing process and appeals to the federal courts is very

uncertain. The first few proposed plants that actually go through the COL process with a plant

that meets the pre-approved Standard Design Specifications will teach us a lot about the duration

and costs of this new licensing process. Indeed, the first plants through this process will provide

valuable information to all potential investors in nuclear plants, but there is no obvious way for

them to appropriate the value of this information to help them to cover any "first of a kind"

regulatory costs they incur by being the first projects that go through the new regulatory process.

12 Nuclear Energy Institute, www.nei.org accessed November 27, 2006. 13 Ibid.

15

Financial incentives in the energy policy act of 2005

I think that it is fair to say that prior to the passage of the Energy Policy Act of 2005 ("the

Act"), investors were not exactly beating down the door of the NRC to file applications of Early

Site Permits or COLs. However, the Act provides a number of significant financial incentives to

the first few plants that enter the COL process, are built and ultimately begin to operate. These

incentives, combined with rising fossil fuel costs, rising wholesale market prices, and growing

recognition that CO2 prices may be imposed at some point within the life of a new plant that

enters construction today, have stimulated much more serious interest among investors in

building new nuclear plants.

The Act provides for a 1.8 cent/kWh investment tax credit for new nuclear capacity

during its first 8 years of operation. This subsidy is limited to no more than $125 million per

year per 1,000 Mw of capacity and no more that 6,000 Mw of new capacity can receive this

subsidy. In addition, new nuclear plants are eligible to apply for loan guarantees for up to 80%

of a plant's construction cost. These loan guarantees will reduce the cost of debt financing for

projects that receive them and allow the financing of the projects to be more highly leveraged.

These subsidies reduce the life-cycle costs of a new nuclear plant by on the order of $20/Mwh,

assuming that they operate with 85% capacity factors (IEA, p. 376).14 The Act also provides

"insurance" against regulatory delays for the first 6,000 Mw of new capacity that applies for a

COL. The first two plants are eligible for up to $500 million of payments for the costs of

regulatory delay and the next four plants for up to $250 million each. The details of how much

in loan guarantees will actually be made available by the federal government (all generating

plants that do not produce greenhouse gases are eligible), how investment tax credits will be

allocated if more than 6,000 Mw of new capacity enters service during the eligibility window

specified in the Act, and how the costs of regulatory delay will be determined are yet to be

specified by the federal government.

The rationale for these financial incentives is that there are "first mover costs," including

potential costs of regulatory delays, that are acting as barriers to restarting construction of new

nuclear plants and allowing the development of new nuclear designs to move down a learning

curve to realize cost savings. According to this theory, by subsidizing the first 6,000 Mw of new

capacity, the direct costs of construction and uncertainty about the speed and costs of the

14 My colleagues at MIT have derived estimates of the aggregate value of the subsidy that are only slightly smaller.

16

regulatory process will fall and nuclear investment will then be economic without further special

subsidies. It is a "learning by doing" type of justification.

Since the passage of the Act, intentions to pursue roughly two-dozen new nuclear plants

have been "announced" by individual generating companies or consortia of generating

companies.15 Many of these companies have substantial nuclear operating experience. However,

announcing the intention to pursue a project and actually making a significant financial

commitment to proceed with a COL and construction once it is granted are not the same thing.

No applications for COLs have yet been made and as far as I know there have been no firm

contracts consummated to purchase equipment and construction services from nuclear equipment

suppliers and construction companies. The first COL applications are forecast for late in 2007

and this suggests that it is unlikely that any new nuclear capacity will enter service much before

2015.

Clearly, a subsidy of about $20/Mwh plus insurance against the costs of regulatory delay

is a significant incentive for nuclear plant investments that are eligible to receive the subsidy. It

is equivalent to placing a price of about $25/ton of CO2 emissions from pulverized coal plants.

However, since the subsidies are available for only the first 6,000 Mw of new nuclear capacity,

the long term effects of the program on investment in nuclear capacity will depend on (a) the

industry being able credibly to reduce construction costs to a level significantly below

$2,300/kW ($2006), (b) to keep construction time at 5 years or less, or (c) for the federal

government to introduce a program that places a price on CO2 emissions of at least $25/ton of

CO2.

These conclusions are consistent with the EIA's forecasts of investment in new nuclear

plants between 2005 and 2030 contained in its 2006 Annual Energy Outlook (IEA AEO 2006,

pp. 79-80). The EIA reference case (which does not place a price on CO2 emissions) reflects the

subsidies included in the Act and projects 6,000 Mw of new nuclear capacity by 2030. The

results for the reference case suggest that these nuclear subsidies make the 6,000 Mw they apply

to economical investments, but that nuclear plant construction costs do not fall fast enough to

make further post-subsidy nuclear plant investment economical.16 If this result is realized in

reality it would suggest that the subsidy program had failed to achieve its primary "learning by

15 Nuclear Energy Institute web site http://www.nei.org accessed November 18, 2006.

17

doing" goals. The EIA's 2006 AEO also presents results for two alternative cases. The first case

reduces the construction costs of new nuclear plants by 20% from the reference case values

(roughly consistent with the low construction cost case in the MIT study). With a 20%

construction cost reduction, about 34,000 Mw of new nuclear capacity enters service by 2030

(IEA AEO 2006, p. 84). I consider this to be a challenging but possible result to achieve.

Finally, the EIA 2006 AEO contains a "vendor cost goal" case that requires an average reduction

in construction costs from the reference case of 35%, falling to a 44% reduction by 2030 (IEA

AEO 2006, p. 84). This case yields 77,000 Mw of new nuclear capacity by 2030. I consider the

cost reduction assumptions in this case to be implausible. However, a cost reduction of 20% plus

a carbon dioxide emissions price of about $25/ton CO2 could yield a similar result.

Before concluding this section, there is one observation that I want to make about

"national security" rationales for nuclear power that appear frequently in the media and are

sometimes exploited by nuclear power proponents. First, almost no oil is used to generate

electricity in the U.S. and this fact is not likely to change in the future. Second, in the EIA AEO

2006 scenarios discussed above, new nuclear plants primarily are substitutes for new coal plants.

As a result, the investment in nuclear capacity does not have a significant impact on imports of

oil or liquefied natural gas (LNG), which is forecast to grow rapidly over this period.

Accordingly, there does not appear to be an "energy security" case for investment in new nuclear

plants. As I have already indicated, I believe that the primary policy rationale for providing

special subsidies to "first mover" nuclear power plants is to maintain nuclear power as an option

for generating electricity that would support a U.S. policy to reduce CO2 emissions that would

otherwise be produced from coal burning power plants.

5. Other Potential Institutional Barriers

There are other potential institutional barriers to a significant growth in new nuclear

capacity that need to be taken into account. First, the unsettled state of electricity sector

restructuring and deregulation (Joskow 2006a) suggests to me that it is unlikely that we will see

much if any investment in new nuclear capacity in states that have adopted competitive market

models, at least until wholesale and retail market designs deal with design flaws that create

general disincentives for investment in new generating capacity (Joskow 2006b). Moreover,

18

stabilizing wholesale and retail market designs so that investors can count on a clear, stable, and

fair market environment will need to occur to support this type of investment. Second, potential

investors in new nuclear plants will still have to deal with state and local regulatory authorities

and potential resistance from anti-nuclear groups that have been quiet during the long hiatus in

new nuclear plant construction. Historically, the greatest opposition to nuclear plants has been in

the Northeast and the West Coast, especially California.17 Many of the states that have adopted

competitive market models also happen to be located in these regions of the country.

The current state of competitive electricity markets in the U.S. combined with local

opposition to new nuclear plants is not an attractive combination for potential investors.

(Ironically, the economics of nuclear investment are quite promising in New England, New

York, New Jersey, California and Texas because existing natural gas fueled generation is on the

margin in many hours of the year and, as a result, wholesale prices are quite high there.) On the

other hand, there are many existing nuclear plant sites, including in states which have not

adopted competitive electricity market models, that were designed to accommodate more nuclear

plants than were actually built on these sites. There is room to build additional nuclear capacity

on several of these sites. In most cases, the residents in the areas near these sites have come to

accept nuclear power plants and, indeed, to enjoy the property tax revenue that they produce.

They are likely to be favorably disposed to investments to expand capacity on at least some of

these existing sites.

The final potential institutional barrier to significant investment in new nuclear capacity

in the U.S. is the continued inability of the federal government to fulfill its commitment to take

back spent nuclear fuel and to provide for its safe long-term storage and disposal. The progress

on getting the Yucca Mountain site licensed to receive spent fuel continues to be subject to

delays and it is opposed by powerful politicians in Nevada. While investment in new nuclear

plants might proceed without a complete resolution of waste disposal issues, unless credible

interim and final waste disposal institutions are placed into operation, it may create significant

investor disincentives and public opposition to a major program of nuclear expansion.

17 Of course it is possible that the current concerns about CO2 emissions will make groups that historically opposed nuclear power more favorably disposed towards it now, and there is some evidence that environmental groups have moved in this direction.

19

6. Conclusions

The future for investment in new nuclear plants in the U.S. is brighter than it has been for

many years. The support and subsidies from the federal government contained in the Energy

Policy Act of 2005 and the earlier reforms in the federal licensing process have given the nuclear

industry what it says it needs to "get over the hump" and stimulate a significant program of

investment in new nuclear power plants. It's now up to the industry to deliver on its claims about

construction costs, operating performance, and financing costs.

All things considered, I believe that investment in new nuclear plants is likely to proceed

more slowly than may be implied by the recent euphoria in the industry. I believe that projects

are most likely to proceed (a) on existing sites, (b) in states that have not adopted competitive

market models, and (c) where there is support from local authorities. Among the competitive

states, Texas is the most likely candidate for investment in new nuclear plants for both economic

and political reasons. A nuclear investment program will be larger and proceed more quickly if

the nuclear equipment vendors and the construction firms are willing to take on more of the

construction cost and operating performance risk than they did during the 1980s, at least until the

first dozen of so plants are completed and credible information about construction costs,

construction time, and regulatory costs and delays have been confirmed by actual experience

rather than hypothetical spreadsheet calculations. Finally, if the U.S. adopts a cap and trade

program for CO2 emissions that yield prices in the range of $25 to $50/ton of CO2, it will make

investment in new nuclear power plants much more attractive financially than it is today,

especially after the 2005 Act's subsidies are exhausted.

REAL NUCLEAR O&M COSTS ($2003 MILLS/KWH)

0

5

10

15

20

25

30

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

$200

3 M

ILLS

/KW

Hnon-fuel O&M fuel total O&M

Source: EIA Electric Power Annual various years

FIGURE 1

U.S. NUCLEAR PLANT CAPACITY FACTORS1973-2005

40

50

60

70

80

90

100

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

%

FIGURE 2

Source: EIA AER 2005

TABLE 1

Comparative Base Load Real Levelized Costs $2002 cents/kWh

Merchant Traditional Regulatory Financing Model18 Model19

Nuclear Cases Base Case ($2000/kW) 6.7 5.2 25% Lower Construction Costs ($1500/kW) 5.5 4.4 Reduce Construction time By 12 months (4 years) 5.3 3.6 Reduce financing costs to Equivalent coal and CCGT 4.2 3.6 Pulverized Coal 4.2 3.5 Gas CCGT Cases Low gas prices ($3.77/Mcf) 3.8 3.6 Moderate gas prices ($4.42/Mcf) 4.1 4.0 High gas prices ($6.72/Mcf) 5.6 5.7

18 MIT (2003), p. 42 and Appendix to Chapter 5. In the merchant case the plant owner bears all construction cost, operating performance and market risk. 19 Author’s calculations. In the traditional regulated case consumers bear the costs or benefits of all “prudent” construction cost overruns, operating performance variations, and changes in market values.

TABLE 2

Effects of CO2 Prices on Real Levelized Cost of Electricity Merchant Financing Model

($2002 cents/kWh)20

$0/ton CO2 $13.50/ton CO2 $27/ton CO2 $54/ton CO2 Pulverized Coal 4.2 5.4 6.6 9.0 CCGT High Gas Price 5.6 6.1 6.7 7.7 Nuclear Cases Base Case 6.7 6.7 6.7 6.7 -25% Construction Cost Case 5.5 5.5 5.5 5.5 Lowest cost Case 4.2 4.2 4.2 4.2

20 See Table 1 for the $0/ton CO2 case. For the other cases see MIT (2003), page 42 and appendix to Chapter 5. Note that carbon prices expressed in $US(2002) per metric ton of carbon in the MIT Report have been express in $US(2002) per metric ton of CO2 here by dividing by 3.67 for easier comparisons with other studies.

References Institute of Nuclear Power Operation (INPO), Annual Report 2005, http://www.nei.org/documents/WANO_Performance_Indicators_2005.pdf, accessed December 1, 2006. International Energy Agency (IEA 2006), World Energy Outlook, November 2006. Joskow, Paul L. (Joskow 2006a), "Markets for Power in the United States: An Interim Assessment," Energy Journal, 27(1), 1-36, 2006. Joskow, Paul L. (Joskow 2006b), "Competitive Electricity Markets and Investment in New Generating Capacity," The New Energy Paradigm, (Dieter Helm, editor), Oxford University Press, forthcoming. http://econ-www.mit.edu/faculty/download_pdf.php?id=1348 Joskow, Paul L. and Richard Schmalensee, "Incentive Regulation for Electric Utilities," Yale Journal on Regulation, December 1986, 1-49. Massachusetts Institute of Technology (MIT 2006)), The Future of Nuclear Power: An Interdisciplinary Study, 2003. http://web.mit.edu/nuclearpower/ United States Energy Information Administration (EIA AEO 2006), Annual Energy Outlook 2006, February 2006 DOE/EIA-0383 (2006). United State Energy Information Administration (EIA AEO 1998), Annual Energy Outlook 1998, December 1997, DEO-EIA-0383(98). United States Energy Information Administration (EIA AER 2005), Annual Energy Review 2005, July 2006, DOE/EIA-0384(2005). United States Energy Information Administration, Electric Power Annual, various years.